Author Archive

Assembled annunciator controller

Sunday, August 17th, 2008

I got my annunciator controller board back from ExpressPCB. Their board layout software is pretty terrible, but the finished product looks okay. If you look closely, you might be able to spot the switch mounting pad I had to grind off because it was too close to the processor reset line. A real board layout tool would have caught that with a design rule check, but of course the crappy free software doesn't have one. Oh well – that turned out to be the only mistake I made (and believe me, nobody is more surprised than I am).

I grabbed a spare D-sub connector and made a little test harness to power the board and test its inputs and outputs.

In keeping with the time-honored traditions of testing out a new PCB design, I soldered the minimal components needed to get the processor running (power supply, oscillator, and the microcontroller itself) and hooked it up to the debugger to see if I could make it blink an LED.

It blinks! This means that the microcontroller is alive and can be programmed. It doesn't look like much, but if a board can get to this stage, then almost everything is working.

Here's the finished product, with voltage dividers, filters, and driver transistors in place. It will be able to sense up to 8 inputs and control up to 6 annunciator lights, depending on how I choose to write the finished firmware. The end goal here is to be able to do a few things with warning lights that are just slightly more sophisticated than simply hooking up a lamp.

Panel planning

Sunday, July 27th, 2008

I bought a copy of DeltaCad and spent the weekend teaching myself to use it well enough to create a design for my instrument panel. Unfortunately I can't put a picture of it here for reasons not worth going into. But it's pretty cool.

Designing a board

Sunday, July 20th, 2008

Your humble author apologizes for the lack of blog-updating. Working sixty hour weeks is incompatible with airplane building! Also, the garage is a furnace this month. However, I did manage to sneak in some time this weekend playing with electronics. I'm designing an annunciator light controller for the panel, which has turned out to be a fun way to exercise the very rusty hardware skills I haven't used since I became a full-time software weenie.

I soldered up a little microcontroller board I bought as a cheap mail-order parts kit:

Then I breadboarded a prototype circuit to verify that the basic design was sound. For this part I raided some old boxes of electronic junk I've been hauling around for at least a dozen years. You just never know when you'll need a 10uF cap on a Saturday night!

Then I spent an afternoon routing a little board. When I used to get paid to design PCBs, I had access to some pretty nice CAD tools. Not so much anymore – the free ExpressPCB software is much less capable, but I got the job done eventually. Next time I might try Eagle, but I think that all the free board layout tools pretty much suck in different ways.

I still haven't settled on the actual annunciator lights I want to use, but at least by next week I should have all the parts I need to make a board to control them.

Ordered propeller

Thursday, July 10th, 2008

Last night I sent Van's a big chunk of change for a propeller – specifically, a Hartzell C2YR-1BFP/F7497-2. That's an aluminum constant-speed prop with a pair of 72" blended-airfoil blades specifically designed for the two-place RV's, and will look something like this.

I kind of wish I'd ordered the prop sooner, since it has a 12-16 week lead time and I find myself with the following chicken-and-egg scenario: In order to finish the FWF plumbing and wiring I kind of need to have the baffles in place, for which I need to have the cowl fitted first, which means I need the prop to fit the cowl. Actually I guess that's a whole series of chickens/eggs. Theoretically you can fit the cowl without the prop if you make some kind of spacer, but even Van's suggests you wait for the real thing to arrive if you have a constant-speed prop like I will. Apparently there's something different about how the spinner is mounted…? We'll see.

Meanwhile, I suppose I will continue to work on interior fuselage systems, when I have time.

Sniffle valve

Sunday, June 29th, 2008

Fuel-injected Lycoming engines can sometimes experience a phenomenon wherein excess fuel drips down the induction tubes and pools in the intake manifold after shutdown. This can be a problem, since it could cause a troublesome fire during the next start. To alleviate this, you install a simple little one-way check valve in the bottom of the intake, which lets the fuel drip out of the engine onto the ground. For whatever reason, this item is colloquially referred to as a sniffle valve.

Here's a bad picture of the one I bought from Airflow Performance for a few bucks – it's basically just a normal AN fitting that's been drilled out and had a ball bearing and retainer clip inserted. Reduced pressure inside the manifold causes the ball to be sucked up to close the valve when the engine is running, but it drops down and lets the fuel drip out when the engine is shut down.

I knew from previous reading that the sniffle valve is too close to the exhaust pipes on an IO-360 with horizontal induction, so I installed it in a 45-degree brass street elbow to clear the pipes. The exact orientation of the valve is said not to matter much, as long as the ball is free to drop when the airplane is at rest. I also attached a 1/4" hose barb to 1/4" AN flare adapter, similar to the one I used for the fuel pump drain line.

I ran another 1/4" aluminum line from the firewall up to the sniffle valve, using a length of 1/4" rubber fuel hose as a flexible coupling. The forward (engine) end of the aluminum tube is ever so slightly flared to give the clamps something to work with – using a real beading tool would have been best, but I can't afford one and this is not a critical application. The aft end of the drain tube is attached to the firewall flange with an adel clamp, just like its twin the fuel drain line.

One more of the million firewall forward tasks is complete… and with that, I'm off to try and make a dent in the beer and ice cream still left over from Saturday's party.